01.12.2012 Views

THE UNIVERSITY OF LEIPZIG

THE UNIVERSITY OF LEIPZIG

THE UNIVERSITY OF LEIPZIG

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

6 Institutional Mapping at the Local Level<br />

6.1 Sample size determination and identification of clusters<br />

A total of 220 farmers were interviewed in a local institutional mapping household survey<br />

conducted in villages around Kakamega Forest. The survey respondents, who were<br />

predominantly farmers, are spread in the different clusters that were statistically constituted<br />

by the district statistical office in Kakamega. These clusters also fall under the sampling frame<br />

that is used by Kenya Central Bureau of Statistics. The usual sampling procedure in Kenya<br />

consists of randomised selection of clusters corresponding with enumeration areas within a<br />

master sample, with a probability equivalent to size in the number of households in every<br />

selected cluster. In each cluster, there are 168 house holds.<br />

Therefore, delectation of clusters was based on the sampling frame developed by the National<br />

Sample Survey Evaluation Programme (NASSEP III). The NASSEP III sampling frame is a<br />

two-stage stratified cluster sample design with individual districts forming the strata. In this<br />

regard, Kakamega District has 26 clusters. From these, 11 clusters were identified for the<br />

survey. These clusters eventually provided the principle sampling points. The Principle<br />

Sampling Point (PSP) was a village commonly known as a location. These clusters were<br />

purposively sampled given their location and proximity to the forest.<br />

Because the survey intended to investigate the role of local notions in institutional framing,<br />

clusters with localities in or around the forest were found to be useful. Hence the areas around<br />

the forest were deliberately represented in the sample. In each cluster, two villages were<br />

visited and in each village, 10 households were randomly chosen and interviewed. Hence,<br />

from each cluster 20 households were interviewed. In total 220 house holds were sampled.<br />

The sample size for the house holds to be covered in the local mapping survey was<br />

statistically determined based on a precision of 5% and a confidence level of 95%. 595<br />

On top of data collected from the individual household respondents, 11 focus group<br />

discussions were held to augment the quantitative data.<br />

Having processed the sample size, the survey team proceeded to identify the eligible<br />

enumeration areas. It is important to point out that each cluster contained 168 house holds.<br />

These clusters are thought to represent the different geographical scope of the district.<br />

However, these same clusters also take into consideration different other aspects such as<br />

semi-urban/rural stratifications, economic and social differentiations as well as livelihood<br />

zonations. 596<br />

Based on the assumption that local notions are more understood and appreciated by household<br />

heads, the interviews were restricted to the two household heads. In the survey, either the<br />

husband or wife was taken as the respondent. The average distance was 4 kilometres from<br />

Kakamega forest. This distance was ascertained by making rough approximation by the<br />

survey enumerators. In the survey the mean number of adults residing in a particular<br />

595 Bartllett, J.E, J.W, Kotrlik and Higgins, C. 2001. Opcit.<br />

596 CBS. 2006. Kenya Economic survey. Nairobi: Government Printer<br />

126

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!